Disk migration RAID 1 problem

[Frank Euskirchen]

Hi,
I want to migrate my Win2K-Installtion from a ML350 G2 with Smart Array 641 to ML 370 G3 with Smart Array 642 Controller. The OS-Partition of the source server is mirrored with RAID1.
The idea is to take off one of the mirrored disks and put it in server 2. After this the 642-controller reports "1787-Slot1 Drive Array Operating in Interim Recovery Mode". So far so good. I then performed a Win2k repair installation and my system is up again. Everything seems to run ok.
As next step I added a second disk to rebuild the mirror. But ACU still tells me that the array is running in Interim Recovery Mode.
I can't find a way to rebuild or to manage the mirror. Also I can't tell the controller to operate with the second disk.
Can anybody help me?
TIA
Frank

[Oleg Koroz]

did you pull drive, on the fly or you powered down server ?

[Frank Euskirchen]

Hi,
I pulled the drive (hot plug) without shutting down the server.
Is there a difference between the two methods?
Thanks
Frank

[Oleg Koroz]

I know some people will say it's safe, ... No
Pulling drive out on the fly can cause all kind Problems: including data loss, Array configuration corruption, RIS ..est
Hot Plug is not a Hot Pull - see attachment

[Frank Euskirchen]

But what happens to RAID1 configurations when the server crashes?
In that situation I don't have the choice to shut down the server to pull out the disk. In other words: I don't have a chance to eject the disk in a "safe mode".
Isn't there any chance to rebuild the mirror?
Does the state stay in "Interim Recovery Mode" or are there any tools to rebuild the mirror?

Frank

[chongkan]

Hi Frank.

Did you changed the drive's position? bay?

Run ADU, save report and call HP support, ask the agent if he can help you finding SCSI backplane errors on the report.

The rebuild process might be been queued by other operation.

If the Raid 1 is still running on ACU, try assigning the drive as a hot spare for the raid to see if the controller start the rebuild on this drive.

Regards.

[chongkan]

Download and install latest firmware before

[Joshua Small_2]

Hi,

With regards to your concerns. I understand the arguement. It "feels" like you should be able to pull a drive out on the fly, because that's exactly what you do when a drive crashes and you hot replace it.

The difference is that when a drive crashes, it's the dead drive you remove. It becomes corrupted by being hot pulled, but it's already gone.

The drive which stays in the server, because there was nothing wrong with it, is unaffected by your actions on the dead drive and hence keeps running.

In other words, a drive is only "consistent" when pulled from a shutdown server, and if you think hard enough, this is not a problem in general.